The present invention relates to an encoding apparatus and an encoding method capable of reducing the overhead on a macroblock basis. A number of reference image decision unit 19 determines that the number of fields of the reference image in encoding is one on the basis of the bit rate of the coded data. If the number of fields of the reference image is determined to be one, a slice header generating unit 13 selects, from among pictures input to an image sorting buffer 12, one of the pictures as the reference image. The present invention is applicable to, for example, encoding apparatuses that perform video coding based on the AVC standard.
Legal claims defining the scope of protection. Each claim is shown in both the original legal language and a plain English translation.
1. An encoding apparatus, comprising: one or more processors configured to: determine that a number of reference pictures used to encode image data is one based on an excess amount by which a file size of the encoded image data exceeds a determined file size; and select data for one image as a reference picture from among the encoded image data based on the determination that the number of reference pictures used is one, wherein the number of reference pictures is determined as one based on the excess amount that is higher than or equal to a first threshold value, wherein the first threshold value is a first minimum value of the excess amount for which the file size of the encoded image data exceeds the determined file size.
An image encoding system reduces computational load by dynamically adjusting the number of reference frames used for encoding. The system checks how much the encoded file size exceeds a target file size. If this "excess amount" is above a threshold, the system switches to using only ONE reference frame for encoding. This threshold is the minimum excess file size at which using a single reference frame becomes beneficial. This reduces the complexity of motion estimation and compensation during encoding, especially when bandwidth or processing power is limited.
2. The encoding apparatus according to claim 1 , wherein the one or more processors are further configured to determine that the number of reference pictures is one based on a bit rate of the encoded image data and based on the file size of the encoded image data that is larger than the determined file size.
Building upon the single reference frame encoding system, the system uses BOTH the excess file size AND the bit rate of the encoded video to determine when to use only one reference frame. If the file size is larger than the target size, AND the bit rate is also high, the system decides to use only one reference frame. Using both metrics provides a more robust decision-making process, ensuring that single-reference frame encoding is only activated when necessary to meet both file size and bit rate constraints.
3. The encoding apparatus according to claim 2 , wherein the one or more processors are further configured to determine that the number of reference pictures is one based on the bit rate of the encoded image data at each point in time that is higher than or equal to a second threshold value, wherein the second threshold value is based on the bit rate that corresponds to the determined file size at the point in time.
Expanding on the previous file size and bitrate method, the system uses the bitrate at EACH point in time during encoding. If the bitrate at any given moment exceeds a dynamically calculated threshold (based on the bitrate needed to meet the target file size at that moment), then the system switches to using only one reference frame. This allows the system to react in real-time to bitrate spikes, optimizing encoding efficiency on a frame-by-frame basis.
4. The encoding apparatus according to claim 1 , wherein the one or more processors are further configured to select the reference picture based on at least one of a motion of a first picture that is encoded or an amount of blur.
In the single-reference frame encoding system, the system selects the specific reference frame based on either the amount of motion within the current frame being encoded OR the amount of blur present in the current frame. By considering these factors, the system can choose the most relevant single reference frame, minimizing the impact on video quality while still reducing the encoding complexity.
5. The encoding apparatus according to claim 4 , wherein the one or more processors are further configured to select one of a second picture with a field type same as that of the first picture or a third picture with a time distance from the first picture shorter than a second threshold value as the reference picture based on the at least one of the motion of the first picture or the amount of blur.
The single reference frame selection considers related pictures. The system selects the reference frame based on either motion or blur in the current frame. It picks either a picture with the SAME field type (top/bottom field in interlaced video) OR a picture that is CLOSEST in time to the current frame (within a defined time threshold). This selection minimizes artifacts by selecting a temporally or structurally similar reference frame when only one reference frame is used.
6. The encoding apparatus according to claim 5 , wherein the one or more processors are further configured to select the third picture as the reference picture based on at least one of an amount of motion of the first picture, a Motion Estimation residual, or a Motion Vector length that is greater than or equal to a third threshold value.
When selecting the single reference frame, if motion is high (motion amount, motion estimation residual, or motion vector length above a threshold), the system prefers the picture CLOSEST in time. High motion suggests objects are moving rapidly, so a temporally close frame is more likely to be a good match, improving encoding efficiency with a single reference.
7. The encoding apparatus according to claim 5 , wherein the one or more processors are further configured to select the second picture with the field type same as that of the first picture as the reference picture based on at least one of an amount of motion of the first picture, a Motion Estimation residual, or a Motion Vector length that is less than or equal to a third threshold value.
Conversely, if motion is LOW (motion amount, motion estimation residual, or motion vector length below a threshold), the system prefers the picture with the SAME field type. Low motion implies a more static scene, where matching the field type is more important for preserving detail and avoiding artifacts in interlaced video.
8. The encoding apparatus according to claim 5 , wherein the one or more processors are further configured to select the third picture as the reference picture based on at least one of an amount of motion of the first picture, a Motion Estimation residual, or a Motion Vector length that is less than a third threshold value and the amount of blur that is greater than or equal to a fourth threshold value.
The system chooses the picture closest in time as the reference frame if the amount of motion is low AND the amount of blur is high. High blur suggests that details are already obscured, so matching the field type is less critical, and temporal proximity becomes more important for capturing the overall scene structure.
9. The encoding apparatus according to claim 5 , wherein the one or more processors are further configured to select the second picture as the reference picture based on at least one of an amount of motion of the first picture, a Motion Estimation residual, or a Motion Vector length that is less than a third threshold value and the amount of blur that is less than a fourth threshold value.
The system chooses the picture with the SAME field type as the reference frame if the amount of motion is low AND the amount of blur is also low. In this scenario, it's more important to match the field type for improved quality because lower blur means details are more defined.
10. The encoding apparatus according to claim 1 , wherein the first threshold value is set to a lowest value from among a plurality of threshold values obtained at all points in time for which the file size of the encoded image data exceeds the determined file size.
In the single reference frame system, the threshold value for the excess file size (the amount by which the encoded file size exceeds the determined file size) is set to the LOWEST value among all threshold values for each point in time that the file size is exceeded. In other words, the system actively searches for the most lenient moment to switch to single-reference frame encoding.
11. The encoding apparatus according to claim 1 , wherein the reference picture precedes a picture that is encoded or succeeds the picture, and wherein a field type of the picture and the reference picture are same.
The single reference picture used for encoding can be either a picture that comes BEFORE the current picture being encoded or a picture that comes AFTER. Regardless of whether the reference picture is in the past or future, it must have the SAME field type as the current picture, ensuring compatibility and reducing artifacts.
12. The encoding apparatus according to claim 1 , wherein the number of reference pictures is determined as one for a period in which the excess amount is larger than or equal to a second minimum value of the excess amount, and wherein the second minimum value comprises a value among a plurality of values of the excess amount obtained at each point in time on a time axis.
The encoding system determines that one reference picture is sufficient for a period of time during which the file size excess remains above a second minimum value. This second minimum value represents one of the various recorded excess values over time. This means the system only switches to single-reference encoding when the file size is consistently over the target for a sustained period, avoiding frequent switching.
13. An encoding method, comprising: in an encoding apparatus: determining that a number of reference pictures used to encode image data is one based on an excess amount by which a file size of the encoded image data exceeds a determined file size; and selecting data for one image as a reference picture from among the encoded image data based on the determination that the number of reference pictures used is one, wherein the number of reference pictures is determined as one based on the excess amount that is higher than or equal to a first threshold value, wherein the first threshold value is a minimum value of the excess amount for which the file size of the encoded image data exceeds the determined file size.
An image encoding method reduces computational load by dynamically adjusting the number of reference frames used for encoding. The system checks how much the encoded file size exceeds a target file size. If this "excess amount" is above a threshold, the system switches to using only ONE reference frame for encoding. This threshold is the minimum excess file size at which using a single reference frame becomes beneficial. This reduces the complexity of motion estimation and compensation during encoding, especially when bandwidth or processing power is limited.
Cooperative Patent Classification codes for this invention. Click any code to explore related patents in that topic.
October 5, 2011
November 21, 2017
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.